1 Cengage Learning. All Rights Reserved. May not be scanned, copied duplicated, or posted to a publicly accessible website, in whole or in part. Slides by John Loucks St. Edward’s University
Feb 25, 2016
1 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Slides by
JohnLoucks
St. Edward’sUniversity
Chapter 5 Advanced Linear Programming
Applications Data Envelopment Analysis
• Comparing performance of one branch to the whole
Revenue Management• Maximize revenue from short-term
demand of a fixed-inventory Portfolio Models and Asset Allocation
• Maximize return from a mix of investments
Game Theory• Compete with another player for a fixed
sum (ie market share)
3 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Data Envelopment Analysis Data envelopment analysis (DEA)
• used to determine the relative operating efficiency of units with the same goals and objectives.
• Branch of a bank, franchise of a restaurant, etc
DEA creates a hypothetical composite• weighted average (W1, W2,…) of existing
units.• Optimal weights determined by the analysis
Goal is to determine E, the efficiency index for one unit• If E < 1, unit is less efficient than the
composite and be deemed relatively inefficient.
• If E = 1, unit is believed to be efficient compared to the rest.
4 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Data Envelopment Analysis The DEA Model
MIN Es.t. OUTPUTS
INPUTSSum of weights = 1E, weights > 0
5 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Data Envelopment Analysis Hospital Administrators at four hospitals want
to examine performance General, University, County, State Inputs:
• Number of FTE nonphysician personnel• Amount spent on supplies• Number of bed-days available
Outputs• Patient days of Medicare Service• Patient days of non-Medicare Service• Number of nurses trained• Number of interns trained
6 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Data Envelopment Analysis
Input
Output
7 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Data Envelopment Analysis Output
8 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Data Envelopment Analysis Define the Decision Variables
E = Fraction of County's input resources required, compared to the composite hospitalwg = Weight applied to General hospitalwu = Weight applied to University Hospitalwc = Weight applied to County Hospitalws = Weight applied to State Hospital
9 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Data Envelopment Analysis Define the Objective FunctionSince our objective is to detect inefficiencies,
Minimize the fraction of County’s input resources required by the composite high school:MIN E
10 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Data Envelopment Analysis Define the Constraints
• Sum of the Weights is 1:• wg + wu + wc + ws = 1
Output Constraints• General form for each output:
• output for composite >= output for county• Output for composite =
• (Output for general * weight for general) +(output for university * weight for university) + (output for county * weight for county) +(output for state * weight for state)
11 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Data Envelopment Analysis
Output Constraint for Medicare Patient Days• 48.14wg + 34.62wu + 36.72wc + 33.16ws
>= 36.72 Output Constraint for non-Medicare Patient
Days• 43.10wg + 27.11wu + 45.98wc + 56.46 ws
>= 45.98 Output constraint for nurses
• 253wg + 148wu + 175wc + 160ws >= 175 Output constraint for Interns
• 41wg + 27wu + 23wc + 84ws >= 23
12 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Data Envelopment Analysis Input Constraints:
• General Form• Input for composite <= input for county * E
• Input for composite• (input for general * weight for general) +
(input for university * weight for university) +(input for county * weight for county) +(input for state * weight for state)
Nonnegativity of variables: E, w1, w2, w3 > 0
13 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Data Envelopment Analysis
Input Constraint for FTE non-Physicians• 285.2wg + 162.3wu + 275.07wc + 210.4ws
<=275.7E Input constraint for supply expense
• 123.8wg + 128.7wu + 348.5wc + 154.1ws <= 348.5E
Input constraint for bed-days• 106.72wg + 64.21wu + 104.1wc + 104.04ws
<= 104.1E
14 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Data Envelopment Analysis Computer Solution
OBJECTIVE FUNCTION VALUE = 0.765 VARIABLE VALUE REDUCED COSTS
E 0.765 0.000 W1 0.000
0.235 W2 0.500
0.000 W3 0.500
0.000
15 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Data Envelopment Analysis Computer Solution (continued)
CONSTRAINT SLACK/SURPLUS DUAL VALUES
1 0.000 -0.235
2 65.000 0.000
3 0.000 -0.001
4 170.000 0.000
5 4.294 0.000
6 0.044 0.000
7 0.000 0.001
16 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Data Envelopment Analysis Conclusion
The output shows that the composite school is made up of equal weights of Lincoln and Washington. Roosevelt is 76.5% efficient compared to this composite school when measured by college admissions (because of the 0 slack on this constraint (#4)). It is less than 76.5% efficient when using measures of SAT scores and high school graduates (there is positive slack in constraints 2 and 3.)
17 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Revenue Management Another LP application is revenue
management. Revenue management involves managing the
short-term demand for a fixed perishable inventory in order to maximize revenue potential.
The methodology was first used to determine how many airline seats to sell at an early-reservation discount fare and many to sell at a full fare.
Application areas now include hotels, apartment rentals, car rentals, cruise lines, and golf courses.
18 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Revenue Management
LeapFrog Airways provides passenger service forIndianapolis, Baltimore, Memphis, Austin, and Tampa.LeapFrog has two WB828 airplanes, one based in Indianapolis and the other in Baltimore. Each morningthe Indianapolis based plane flies to Austin with a stopover in Memphis. The Baltimore based plane flies toTampa with a stopover in Memphis. Both planes have a coach section with a 120-seat capacity.
19 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
LeapFrog uses two fare classes: a discount fare Dclass and a full fare F class. Leapfrog’s products, eachreferred to as an origin destination itinerary fare (ODIF), are listed on the next slide with their fares andforecasted demand. LeapFrog wants to determine how many seats it should allocate to each ODIF.
Revenue Management
20 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
ODIF123456789
10111213141516
OriginIndianapoli
sIndianapoli
sIndianapoli
sIndianapoli
sIndianapoli
sIndianapoli
sBaltimoreBaltimoreBaltimoreBaltimoreBaltimoreBaltimoreMemphisMemphisMemphisMemphis
DestinationMemphis
AustinTampa
MemphisAustinTampa
MemphisAustinTampa
MemphisAustinTampaAustinTampa AustinTampa
FareClass
DDDFFFDDDFFFDDFF
ODIFCodeIMDIADITDIMFIAFITF
BMDBADBTDBMFBAFBTFMADMTDMAFMTF
Fare175275285395425475185315290385525490190180310295
Demand
44254015108
26504212169
58481411
Revenue Management
21 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Revenue Management Define the Decision VariablesThere are 16 variables, one for each ODIF:IMD = number of seats allocated to Indianapolis-Memphis-
Discount classIAD = number of seats allocated to Indianapolis-Austin- Discount classITD = number of seats allocated to Indianapolis-Tampa- Discount classIMF = number of seats allocated to Indianapolis-Memphis- Full Fare classIAF = number of seats allocated to Indianapolis-Austin-Full Fare class
22 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Revenue Management Define the Decision Variables (continued)ITF = number of seats allocated to Indianapolis-Tampa- Full Fare classBMD = number of seats allocated to Baltimore-Memphis- Discount classBAD = number of seats allocated to Baltimore-Austin- Discount classBTD = number of seats allocated to Baltimore-Tampa- Discount classBMF = number of seats allocated to Baltimore-Memphis- Full Fare classBAF = number of seats allocated to Baltimore-Austin- Full Fare class
23 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Revenue Management Define the Decision Variables (continued)BTF = number of seats allocated to Baltimore-Tampa- Full Fare classMAD = number of seats allocated to Memphis-Austin- Discount classMTD = number of seats allocated to Memphis-Tampa- Discount classMAF = number of seats allocated to Memphis-Austin- Full Fare classMTF = number of seats allocated to Memphis-Tampa- Full Fare class
24 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Revenue Management Define the Objective Function
Maximize total revenue: Max (fare per seat for each ODIF) x (number of seats allocated to the
ODIF) Max 175IMD + 275IAD + 285ITD + 395IMF + 425IAF + 475ITF + 185BMD + 315BAD + 290BTD + 385BMF + 525BAF +
490BTF + 190MAD + 180MTD + 310MAF +
295MTF
25 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Revenue Management Define the Constraints
There are 4 capacity constraints, one for each flight leg:
Indianapolis-Memphis leg (1) IMD + IAD + ITD + IMF + IAF + ITF < 120
Baltimore-Memphis leg (2) BMD + BAD + BTD + BMF + BAF + BTF
< 120 Memphis-Austin leg (3) IAD + IAF + BAD + BAF + MAD + MAF <
120 Memphis-Tampa leg (4) ITD + ITF + BTD + BTF + MTD + MTF <
120
26 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Revenue Management Define the Constraints (continued)
There are 16 demand constraints, one for each ODIF:
(5) IMD < 44 (11) BMD < 26 (17) MAD < 5
(6) IAD < 25 (12) BAD < 50 (18) MTD < 48
(7) ITD < 40 (13) BTD < 42 (19) MAF < 14
(8) IMF < 15 (14) BMF < 12 (20) MTF < 11
(9) IAF < 10 (15) BAF < 16 (10) ITF < 8 (16) BTF < 9
27 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Revenue Management Computer Solution Objective Function Value = 94735.000
Variable Value Reduced Cost
IMD 44.000 0.000 IAD 3.000 0.000 ITD 40.000
0.000 IMF 15.000
0.000 IAF 10.000
0.000 ITF 8.000
0.000 BMD 26.000
0.000 BAD 50.000
0.000
28 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Revenue Management Computer Solution (continued)
Variable Value Reduced Cost
BTD 7.000 0.000
BMF 12.000 0.000
BAF 16.000 0.000
BTF 9.000 0.000
MAD 27.000 0.000
MTD 45.000 0.000
MAF 14.000 0.000
MTF 11.000 0.000
29 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Portfolio Models and Asset Management Asset allocation involves determining how to
allocate investment funds across a variety of asset classes such as stocks, bonds, mutual funds, real estate.
Portfolio models are used to determine percentage of funds that should be made in each asset class.
The goal is to create a portfolio that provides the best balance between risk and return.
30 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
John Sweeney is an investment advisor who isattempting to construct an "optimal portfolio" for aclient who has $400,000 cash to invest. There are tendifferent investments, falling into four broad categories that John and his client have identified as potential candidate for this portfolio. The investments and their important characteristics are listed in the table on the next slide. Note that Unidyde Corp. under Equities and Unidyde Corp. under Debt are two separate investments, whereas First General REIT is a single investment that is considered both an equities and a real estate investment.
Portfolio Model
31 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Portfolio Model Exp. Annual
After Tax Liquidity RiskCategory Investment Return Factor FactorEquities Unidyde Corp. 15.0% 100 60(Stocks) CC’s Restaurants 17.0% 100 70 First General REIT 17.5% 100 75Debt Metropolis Electric 11.8% 95 20(Bonds) Unidyde Corp. 12.2% 92 30 Lewisville Transit 12.0% 79 22Real Estate Realty Partners 22.0% 0 50 First General REIT ( --- See above --- )Money T-Bill Account 9.6% 80 0 Money Mkt. Fund 10.5% 100 10 Saver's Certificate 12.6% 0 0
32 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Portfolio Model
Formulate a linear programming problem toaccomplish John's objective as an investment advisor which is to construct a portfolio that maximizes hisclient's total expected after-tax return over the next year, subject to the limitations placed upon him by the client for the portfolio. (Limitations listed on next two slides.)
33 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Portfolio Model
Portfolio Limitations 1. The weighted average liquidity factor for the portfolio must to be at least 65. 2. The weighted average risk factor for the portfolio must be no greater than 55. 3. No more than $60,000 is to be invested in Unidyde stocks or bonds. 4. No more than 40% of the investment can be in any one category except the money category. 5. No more than 20% of the total investment can be in any one investment except the money market fund.
continued
34 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Portfolio Model
Portfolio Limitations (continued) 6. At least $1,000 must be invested in the Money Market fund. 7. The maximum investment in Saver's Certificates is $15,000. 8. The minimum investment desired for debt is $90,000. 9. At least $10,000 must be placed in a T-Bill account.
35 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Portfolio Model Define the Decision Variables
X1 = $ amount invested in Unidyde Corp. (Equities)X2 = $ amount invested in CC’s RestaurantsX3 = $ amount invested in First General REITX4 = $ amount invested in Metropolis ElectricX5 = $ amount invested in Unidyde Corp. (Debt) X6 = $ amount invested in Lewisville TransitX7 = $ amount invested in Realty PartnersX8 = $ amount invested in T-Bill Account X9 = $ amount invested in Money Mkt. Fund
X10 = $ amount invested in Saver's Certificate
36 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Portfolio Model Define the Objective Function
Maximize the total expected after-tax return over the next year:Max .15X1 + .17X2 + .175X3 + .118X4 + .122X5
+ .12X6 + .22X7 + .096X8 + .105X9 + .126X10
37 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Portfolio Model
Total funds invested must not exceed $400,000:(1) X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9 +
X10 = 400,000Weighted average liquidity factor must to be at
least 65:(2) 100X1 + 100X2 + 100X3 + 95X4 + 92X5 + 79X6 +
80X8 + 100X9 > 65(X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9
+ X10)Weighted average risk factor must be no greater
than 55:(3) 60X1 + 70X2 + 75X3 + 20X4 + 30X5 + 22X6 + 50X7
+ 10X9 < 55(X1 + X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9
+ X10)No more than $60,000 to be invested in Unidyde
Corp:(4) X1 + X5 < 60,000
Define the Constraints
38 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Portfolio Model Define the Constraints
(continued)No more than 40% of the $400,000 investment can be
in any one category except the money category:(5) X1 + X2 + X3 < 160,000(6) X4 + X5 + X6 < 160,000(7) X3 + X7 < 160,000No more than 20% of the $400,000 investment
can bein any one investment except the money market
fund:(8) X2 < 80,000 (12) X7 < 80,000(9) X3 < 80,000 (13) X8 < 80,000(10) X4 < 80,000 (14) X10 < 80,000(11) X6 < 80,000
39 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Portfolio Model Define the Constraints
(continued)At least $1,000 must be invested in the Money Market fund:
(15) X9 > 1,000
The maximum investment in Saver's Certificates is $15,000:
(16) X10 < 15,000 The minimum investment the Debt category is
$90,000:(17) X4 + X5 + X6 > 90,000 At least $10,000 must be placed in a T-Bill account:(18) X8 > 10,000Non-negativity of variables: Xj > 0 j = 1, . . . , 10
40 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Portfolio Model Solution Summary
Total Expected After-Tax Return = $64,355X1 = $0 invested in Unidyde Corp. (Equities)X2 = $80,000 invested in CC’s RestaurantsX3 = $80,000 invested in First General REITX4 = $0 invested in Metropolis ElectricX5 = $60,000 invested in Unidyde Corp. (Debt) X6 = $74,000 invested in Lewisville TransitX7 = $80,000 invested in Realty PartnersX8 = $10,000 invested in T-Bill Account X9 = $1,000 invested in Money Mkt. Fund
X10 = $15,000 invested in Saver's Certificate
41 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Introduction to Game Theory In decision analysis, a single decision maker
seeks to select an optimal alternative. In game theory, there are two or more decision
makers, called players, who compete as adversaries against each other.
It is assumed that each player has the same information and will select the strategy that provides the best possible outcome from his point of view.
Each player selects a strategy independently without knowing in advance the strategy of the other player(s).
continue
42 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Introduction to Game Theory The combination of the competing strategies
provides the value of the game to the players. Examples of competing players are teams,
armies, companies, political candidates, and contract bidders.
43 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Two-person means there are two competing players in the game.
Zero-sum means the gain (or loss) for one player is equal to the corresponding loss (or gain) for the other player.
The gain and loss balance out so that there is a zero-sum for the game.
What one player wins, the other player loses.
Two-Person Zero-Sum Game
44 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Competing for Vehicle SalesSuppose that there are only two vehicle
dealer-ships in a small city. Each dealership is consideringthree strategies that are designed to take sales of new vehicles from the other dealership over afour-month period. The strategies, assumed to be the same for both dealerships, are on the next slide.
Two-Person Zero-Sum Game Example
45 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Strategy Choices Strategy 1: Offer a cash rebate
on a new vehicle. Strategy 2: Offer free optional
equipment on a new vehicle.
Strategy 3: Offer a 0% loan on a new vehicle.
Two-Person Zero-Sum Game Example
46 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
2 2 1
CashRebate
b1
0%Loan
b3
FreeOptions
b2
Dealership B
Payoff Table: Number of Vehicle Sales Gained Per Week by
Dealership A (or Lost Per Week by
Dealership B)
-3 3 -1 3 -2 0
Cash Rebate a1
Free Options a2
0% Loan a3
Dealership A
Two-Person Zero-Sum Game Example
47 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Step 1: Identify the minimum payoff for each row (for Player A).
Step 2: For Player A, select the strategy that provides
the maximum of the row minimums (called
the maximin).
Two-Person Zero-Sum Game
48 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Identifying Maximin and Best Strategy
RowMinimum
1-3-2
2 2 1
CashRebate
b1
0%Loan
b3
FreeOptions
b2
Dealership B
-3 3 -1 3 -2 0
Cash Rebate a1
Free Options a2
0% Loan a3
Dealership A
Best Strategy
For Player AMaximinPayoff
Two-Person Zero-Sum Game Example
49 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Step 3: Identify the maximum payoff for each column
(for Player B). Step 4: For Player B, select the strategy that
provides the minimum of the column
maximums (called the minimax).
Two-Person Zero-Sum Game
50 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Identifying Minimax and Best Strategy
2 2 1
CashRebate
b1
0%Loan
b3
FreeOptions
b2
Dealership B
-3 3 -1 3 -2 0
Cash Rebate a1
Free Options a2
0% Loan a3
Dealership A
Column Maximum 3 3 1
Best Strategy
For Player B
MinimaxPayoff
Two-Person Zero-Sum Game Example
51 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Pure Strategy
Whenever an optimal pure strategy exists: the maximum of the row minimums equals
the minimum of the column maximums (Player A’s maximin equals Player B’s minimax)
the game is said to have a saddle point (the intersection of the optimal strategies)
the value of the saddle point is the value of the game
neither player can improve his/her outcome by changing strategies even if he/she learns in advance the opponent’s strategy
52 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
RowMinimum
1-3-2
CashRebate
b1
0%Loan
b3
FreeOptions
b2
Dealership B
-3 3 -1 3 -2 0
Cash Rebate a1
Free Options a2
0% Loan a3
Dealership A
Column Maximum 3 3 1
Pure Strategy Example
Saddle Point and Value of the Game
2 2 1
SaddlePoint
Value of thegame is 1
53 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Pure Strategy Example
Pure Strategy Summary Player A should choose Strategy a1 (offer a
cash rebate). Player A can expect a gain of at least 1
vehicle sale per week. Player B should choose Strategy b3 (offer a
0% loan). Player B can expect a loss of no more than
1 vehicle sale per week.
54 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Mixed Strategy
If the maximin value for Player A does not equal the minimax value for Player B, then a pure strategy is not optimal for the game.
In this case, a mixed strategy is best. With a mixed strategy, each player employs
more than one strategy. Each player should use one strategy some of
the time and other strategies the rest of the time.
The optimal solution is the relative frequencies with which each player should use his possible strategies.
55 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Competing for Vehicle Sales Let us continue with the two-dealership
gamepresented earlier, but with a change to one payoff.If both Dealership A and Dealership B choose to offer a 0% loan, the payoff to Dealership A is nowan increase of 3 vehicle Sales per week. (The revised payoff table appears on the next slide.)
Two-Person Zero-Sum Game Example #2
56 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Column Maximum
2 2 11
CashRebate
b1
0%Loan
b3
FreeOptions
b2
Dealership B
Payoff Table: Number of Vehicle Sales Gained Per Week by
Dealership A (or Lost Per Week by
Dealership B)
-3 3 -1-3 3 -2 3
-2
Cash Rebate a1
Free Options a2
0% Loan a3
Dealership A
Two-Person Zero-Sum Game Example #2
3 3 3
Row min
57 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Let us first consider the game from the point of view of Dealership A.
Dealership A will select one of its three strategies based on the following probabilities:PA1 = the probability that Dealership A selects strategy a1
PA2 = the probability that Dealership A selects strategy a2 PA3 = the probability that Dealership A selects strategy a3
Two-Person Zero-Sum Game Example #2
58 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Two-Person Zero-Sum Game Example #2 The expected gain under a specific competitor’s
strategy is a weighted average of: The gain under each of your strategies, times… The probability that you will choose said
strategy… Repeated For each of your competitor’s strategy
Dealership B Strategy Expected Gain for Dealership A
b1 EG(b1) = 2PA1 – 3PA2 + 3PA3
b2 EG(b2) = 2PA1 + 3PA2 – 2PA3 b3 EG(b3) = 1PA1 – 1PA2 + 3PA3
59 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Two-Person Zero-Sum Game Example #2 Define GAINA to be the optimal expected gain in
vehicle sales for Dealership A, which we want to maximize.
Thus, the individual expected gains, EG(b1), EG(b2) and EG(b3) must all be greater than or equal to GAINA.
For example,2PA1 – 3PA2 + 3PA3 > GAINA
Also, the sum of Dealership A’s mixed strategy probabilities must equal 1.
This results in the LP formulation on the next slide …..
60 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Dealership A’s Linear Programming FormulationMax GAINA
s.t.2PA1 – 3PA2 + 3PA3 – GAINA > 0
(Strategy b1)2PA1 + 3PA2 – 2PA3 – GAINA > 0 (Strategy
b2) 1PA1 – 1PA2 + 0PA3 – GAINA > 0 (Strategy b3)
PA1 + PA2 + PA3 = 1 (Prob’s sum to 1)
PA1, PA2, PA3, GAINA > 0 (Non-negativity)
Two-Person Zero-Sum Game Example #2
61 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Computer Solution: Dealership AOBJECTIVE FUNCTION VALUE = 1.333
VARIABLE VALUE REDUCED COSTS PA1 (cash) 0.833
0.000 PA2 (opts) 0.000 1.000 PA3 (0%) 0.167
0.000 GAINA 1.333 0.000
Two-Person Zero-Sum Game Example #2
62 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Computer Solution: Dealership ACONSTRAINT SLACK/SURPLUS DUAL VALUES
1 0.833 0.000
2 0.000 -0.333
3 0.000 -0.667
4 0.000 1.333
Two-Person Zero-Sum Game Example #2
63 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Dealership A’s Optimal Mixed Strategy• Offer a cash rebate (a1) with a probability of
0.833• Do not offer free optional equipment (a2)• Offer a 0% loan (a3) with a probability of 0.167
The expected value of this mixed strategy is a gain of
1.333 vehicle sales per week for Dealership A.
Two-Person Zero-Sum Game Example #2
64 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Let us now consider the game from the point of view of Dealership B.
Dealership B will select one of its three strategies based on the following probabilities:PB1 = the probability that Dealership B selects strategy b1
PB2 = the probability that Dealership B selects strategy b2 PB3 = the probability that Dealership B selects strategy b3
Two-Person Zero-Sum Game Example #2
65 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Because the numbers are reversed, we calculate the Expected Loss, instead of Expected GainDealership A Strategy Expected Loss for Dealership B
a1 EL(a1) = 2PB1 + 2PB2 + 1PB3
a2 EL(a2) = -3PB1 + 3PB2 – 1PB3 a3 EL(a3) = 3PB1 – 2PB2 + 3PB3
Two-Person Zero-Sum Game Example #2
66 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Two-Person Zero-Sum Game Example #2 Define LOSSB to be the optimal expected loss in
vehicle sales for Dealership B, which we want to MINIMIZE.
Thus, the individual expected losses, EL(a1), EL(a2) and EL(a3) must all be less than or equal to LOSSB.
For example,2PA1 + 2PA2 + 1PA3 < LOSSB
Also, the sum of Dealership B’s mixed strategy probabilities must equal 1.
This results in the LP formulation on the next slide …..
67 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Dealership B’s Linear Programming FormulationMin LOSSB
s.t. 2PB1 + 2PB2 + 1PB3 – LOSSB < 0
(Strategy a1)-3PB1 + 3PB2 – 1PB3 – LOSSB < 0
(Strategy a2) 3PB1 – 2PB2 + 3PB3 – LOSSB < 0 (Strategy a3)
PB1 + PB2 + PB3 = 1 (Prob’s sum to 1)
PB1, PB2, PB3, LOSSB > 0 (Non-negativity)
Two-Person Zero-Sum Game Example #2
68 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Computer Solution: Dealership BOBJECTIVE FUNCTION VALUE = 1.333
VARIABLE VALUE REDUCED COSTS PB1 0.000 0.833
PB2 0.333 0.000 PB3 0.667 0.000 LOSSB 1.333 0.000
Two-Person Zero-Sum Game Example #2
69 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Computer Solution: Dealership BCONSTRAINT SLACK/SURPLUS DUAL VALUES
1 0.000 0.833
2 1.000 0.000
3 0.000 0.167
4 0.000 -1.333
Two-Person Zero-Sum Game Example #2
70 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Dealership B’s Optimal Mixed Strategy• Do not offer a cash rebate (b1)• Offer free optional equipment (b2) with a
probability of 0.333• Offer a 0% loan (b3) with a probability of 0.667
The expected payoff of this mixed strategy is a loss of1.333 vehicle sales per week for Dealership B.
Note that expected loss for Dealership B is the same asthe expected gain for Dealership A. (There is a
zero-sum for the expected payoffs.)
Two-Person Zero-Sum Game Example #2
71 Slide
© 2011 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
End of Chapter 5